Haitao Yan

Papers

Total Citations

H-Index

About

Haitao Yan is a rising researcher at the intersection of computer vision, human-robot interaction, and robotic automation. His work focuses on enabling intelligent systems to better perceive, predict, and learn from human motion. Yan’s most notable contribution is a novel method for **forecasting 3D whole-body human poses with grasping objects**, a critical capability for safe and intuitive human-robot collaboration. This work, already garnering 6 citations since 2024, addresses a key limitation in existing models by predicting not just major body joints but the full articulated body in interaction with objects. Additionally, Yan has advanced **telerobotic skill transfer** with an unsupervised approach for segmenting complex, high-dimensional robot trajectories. This method, cited 3 times, offers a scalable solution for automating skill training by breaking down intricate teleoperated motions into meaningful segments without requiring labeled data. Through these contributions, Yan is helping to bridge the gap between raw human motion data and the robust, predictive models needed for next-generation collaborative robots and autonomous systems.